pulumipulumiStack: operations-Babylon-4 info: panic: runtime error: invalid memory address or nil pointer dereference [signal 0xc0000005 code=0x0 addr=0x0 pc=0x128929d]

Hi how can I make a fieldgroup as collapsible I tried using the property it’s not working in suitelet

Used the new

factoryFunc

of

aws.serverless.Function

. Not sure if the global approach is the best but it’s really cool to have that

Completely unrelated : if I want to reference a config value inside one of my lambda, I’m forced to use env variables ?

just got pulumi to create a new AKS cluster... deploy a service... create an ip... get a lets-encrypt certificate and it all works. pulumi's awesome!

am i being a silly billy... helm.v2.ChartOpts only takes string not Input<string> how can i specify it from a resource it depends on?

i'll time myself but i think it will take about 1 hr

How should AWS helper functions, like

aws.getCallerIdentity()

, that return a Promise instead of an Output be used?

@full-dress-10026 you can

await

them

Or you can use a Promise as a

pulumi.Input

IIRC

Alternatively, if you're just passing them as resource inputs--yeah

You only need to

await

if you need to use the value in control flow

^^what he said. @stocky-spoon-28903 beat me to it 🙂

You can see an example of awaiting one of these in the snippet I shared above:

let serviceVpcId = service.ecsService.networkConfiguration.apply(async (networkConfig) => {
    let subnetId = networkConfig["subnets"][0];
    let subnet = await aws.ec2.getSubnet({ id: subnetId });
    return subnet.vpcId;
});

Right, that one makes sense. Just await'ing won't cause a potential race condition though?

getCallerIdentity()

is used independently -- not in the

apply

of an output.

You typically want to combine these Promises into your data flow somehow - either constructing another Promise from them that gets passed into an input somewhere, or combining them into other

Outputs

. But you can also use them directly for control flow if you want (though that can limit parallelism and lead to slightly more awkward program structure). Depends a bit on the specific task you are trying to solve for.

I see. I really only need the account id for the caller, and because promises can be used as inputs, this should work:

let awsAccountId = aws.getCallerIdentity().then((value => {
    return value.accountId;
}));

Is there a class to accept a VPC endpoint connection? i.e. https://docs.aws.amazon.com/cli/latest/reference/ec2/accept-vpc-endpoint-connections.html

I cannot see a way to accept a VPC endpoint connection with Terraform either.

Does anyone have an example of constructing a Kinesis Firehose?

Is it ok for the

apply

function to do IO?

Specifically, send an HTTP request.

Yep - totally fine to do I/O in

apply

. Just worth understanding what/when that will run, and whether that matches expectations. During

update

it will always run (assuming resources successfully create). During a

preview

these will only run if the inputs are known (which will only sometimes be the case).

@white-balloon-205 is there an escape hatch for not performing it if you are in the case of a preview with known inputs?

I don’t have that use case right now but may be useful for future

I think you can check runtime.dryRun() to see if your in a preview

That's right -

pulumi.runtime.isDryRun

is available to manually test. Shouldn't normally be needed - but does allow control when it is necessary.

Hmm I could probably make that work but it's not ideal. Is there a way to run some code that only executes after a resource has been created (not updated or deleted)?

Is it possible, either with

serverless

or

aws

to write a lambda function inline, in JS? The goal is to bind it to Kinesis, but I'd like to not have to create a separate folder, if possible. Can you write regular JS here?

let lambda = new aws.lambda.Function("mylambda", {
    runtime: foo,
    code: CAN CODE GO HERE?,
    timeout: 300,
    handler: foo,
    role: role.arn,
    environment: { 
        variables: {
            "COUNTER_TABLE": counterTable.name
        }
    },
}, { dependsOn: [fullAccess] });

Is it possible to run a one-off task (

cloud.Task#run

) on creating a resource? Or is this straying too far away from the scope of pulumi?